Startup Business Models: The Main Types Explained

A startup business model is the logic of how a young company creates value for customers, delivers it, and earns money from it. It's not the product, and it's not the price list. It's the whole arrangement: who the customers are, what they get, who pays, what it costs to serve them and what's left over.

The phrase has a messy history. During the dot-com boom, Joan Magretta wrote in Harvard Business Review, "business model" was a buzzword invoked to glorify half-baked plans, and many people who bought the fantasy got burned. The concept recovered because it names something real. Two companies can sell similar things and run very different businesses, depending on who pays them and how.

This article is the overview for the business-model articles in this collection. It defines the term, shows the questions every model has to answer, walks through six common startup model types, compares them in a table, and points you to a deeper article on each. It's a reference guide, not a recommendation. Which model fits depends on the market, the product and the founders, and practices vary across industries and countries.

What a business model is (and isn't)

Researchers have defined the term in slightly different ways, but the definitions overlap. David Teece's often-cited 2010 paper says that a business model describes the design or architecture of how an enterprise creates, delivers and captures value. He adds that its essence is defining how the enterprise delivers value to customers, gets customers to pay for it, and converts those payments into profit. Teece also describes it as management's hypothesis about what customers want and how the company can organize to meet those needs profitably.

That last word, hypothesis, is useful for founders. A business model isn't a fact about your company. It's a bet, and you test it. The Lean Startup method and the riskiest assumption test exist because early models are mostly guesses.

Zott and Amit offer a second view. They describe a business model as a system of interdependent activities that spans the firm's boundaries, which lets the firm, together with partners, create value and capture a share of it. The point: a model isn't only what your company does. It also covers what partners, suppliers, sellers or customers do.

Two things a business model is not:

• Not a strategy. Strategy is about how you win against competitors. A business model is about how the business works at all. Teece's paper explores how the two connect, and they're related but not identical.
• Not a revenue model. A revenue model is one part of the picture: how money comes in (subscriptions, fees, ads, licenses). The business model also covers customers, costs, delivery and partners.

The questions every model has to answer

Alexander Osterwalder and Yves Pigneur's Business Model Canvas is the most widely used way to lay a model out. According to Strategyzer, the canvas's nine blocks cover who you serve, what you offer, how you deliver it, what it costs and what it earns. It also notes that the right side of the canvas is the market, the left side is what it takes to serve that market, and the value propositions sit between them. The Business Model Canvas article covers each block in detail.

You don't need the canvas to see the underlying questions. Whatever the model type, a startup needs an answer to each of these:

1. Who is the customer? Sometimes there are several kinds, and they're not the same as the user.
2. What value do they get? The problem solved or the job done.
3. Who pays, and for what? The payer and the thing being paid for can be different.
4. What is the unit of revenue? A transaction, a subscription seat, a device, a project, a license.
5. What does it cost to deliver one more unit? This drives margin and scaling.
6. What has to be true for it to work? Network size, partner reach, community, manufacturing, standard process.

Question 3 is where the model types below separate most clearly. And question 5 is where the unit economics of the model show up: whether each additional customer, sale or project makes or loses money.

Six common startup business model types

These aren't the only models, and real companies mix them. But they cover a lot of the ground early-stage founders meet, and each has its own structure, risks and metrics. For each one, here's the short version and where to read more.

Marketplace

A marketplace connects buyers and sellers and lets them transact with each other. The operator usually doesn't make the product or hold the inventory. It builds the place where the match happens, sets rules, provides trust tools and often payments, and takes a cut. That cut is typically a commission or fee on each transaction, sometimes combined with listing fees, seller subscriptions or paid promotion.

The defining difficulty is that a marketplace needs both sides before it's useful to either, the chicken-and-egg problem that the economics of two-sided markets describes. Founders usually solve this by starting in a narrow niche. The other persistent risk is leakage: once a buyer and seller trust each other, they may move repeat deals off the platform to avoid the fee. The marketplace business model article covers marketplace types, take rate, GMV versus revenue and liquidity.

Platform

A platform enables interactions between producers and consumers and earns from those interactions. The term overlaps with marketplace, but it's wider. A marketplace is one kind of platform, a transaction platform. Others are innovation platforms, where a base technology lets third parties build their own products on top, and hybrids that combine the two.

Platforms can compound through network effects: each new participant makes the platform more valuable to the others. But they're hard to start and easy to overestimate. Winner-take-all outcomes are a tendency in some markets, not a law. Read more in the platform business model article.

B2B2C

In a B2B2C model, a startup sells to a partner business, and that partner uses the product to serve its own consumers. The startup reaches end users through the partner instead of acquiring each one itself. Typical structures include wholesale or license fees, revenue share, per-member fees and referral commissions.

The appeal is distribution and borrowed trust. The cost is dependence: when one or two partners supply most of the users and revenue, they hold leverage over pricing and terms. Questions of data ownership and brand dilution follow. See the B2B2C business model article for variants and risks.

Open source

An open-source company gives away software whose license meets the Open Source Definition, then earns from what surrounds the code. Common routes are support and subscriptions, an open core with proprietary paid features, a hosted cloud service, and commercial licenses for companies that want to embed the code. Most companies combine several.

The central tension is between community and revenue. Give away too little and the community won't form. Give away too much and nobody needs to pay. Since 2021, a number of well-known companies have moved to source-available licenses, which are not the same as open source. The open-source business model article explains the license differences and monetization options.

Hardware

A hardware startup sells a physical device. That changes the economics, because every unit has a real cost, inventory and tooling tie up cash before revenue arrives, and many products need certification before they can be sold. Common models include a one-time sale, consumables, a device plus a subscription, hardware-as-a-service and B2B or industrial sales.

Many hardware companies attach recurring services to the device because the device itself carries lower margins than software. Hardware iteration is slower and more expensive too, which is why investors tend to view it as harder. The hardware startup business model article covers the development path, certification and cash flow.

Productized service

A productized service sells expertise the way a product is sold: a defined scope, a fixed price, a standard process and a stated timeline. Think of the difference between "we'll scope it after a discovery call" and a package with a set number of deliverables and one price.

It's an appealing starting model because a founder with a skill can sell it with little capital. But people still do the work, so capacity grows with headcount, and scope creep and founder dependence are the common failure points. Some productized services turn into software later; others are better left as well-run services businesses. See the productized service article.

The six types compared

The table below is a simplification. Real companies blend models, and the margin descriptions are directional, not benchmarks.

Two patterns stand out. First, the "who pays" column varies a lot, and misjudging it is one of the most common early mistakes. In a marketplace, the side that benefits most may not be the side you can charge. In B2B2C, the person who uses the product isn't the person who signs the contract. Second, the constraint column tells you what to test first. A marketplace founder tests whether both sides will show up. A hardware founder tests whether the product can be built and sold at a cost that leaves a margin.

Revenue models inside the business models

A business model type isn't the same as how you charge. Within any of the six, a company chooses one or more revenue models. The usual building blocks:

Transaction fees, subscriptions, usage-based pricing, licenses, one-time sales, advertising and freemium are the usual building blocks.

Mixed models are the norm: a marketplace may add seller subscriptions, and an open-source project may sell hosting.

Business model innovation: a new model, not just a new product

Sometimes the model itself is the breakthrough. Johnson, Christensen and Kagermann describe Apple's 2003 launch of the iPod with the iTunes store as an example. In their account, the iPod and iTunes combination became a nearly $10 billion product in three years, close to 50% of Apple's revenue, and Apple's market capitalization rose from about $1 billion in early 2003 to over $150 billion by late 2007.

The lesson for a startup isn't to copy Apple. It's that the pairing of a device with a content store was a model decision, not only a product decision. When you change who pays, what they pay for or how it is delivered, you change the model.

How to choose a model

There's no formula, but a few filters help.

Start from the customer's problem, not the model. Models are containers. If the problem is "small suppliers can't find buyers," a marketplace may fit. If it's "my team repeats the same service work," a productized service may. Picking a fashionable model and searching for a problem is how startups end up with a structure that doesn't match demand. This is why customer discovery comes first.

Check capital and time. Hardware needs money before the first sale, a marketplace may need to subsidize one side, and a productized service can start with almost nothing.

Ask what must be true, then test the riskiest thing. For a platform it may be that participants will join without a network. For B2B2C, that a partner will actually promote you. For open source, that someone will pay for what's not free. Then run the cheapest experiment you can.

Check the unit economics early. Compute what one customer, transaction or project earns after the cost of serving it.

Expect to change it. A change large enough to alter who pays or what is sold is often called a pivot.

Key Facts

• A business model describes how a company creates, delivers and captures value. Teece's 2010 paper says its essence is defining how the enterprise delivers value to customers, gets them to pay and converts payments into profit.
• Zott and Amit conceptualize it as a system of interdependent activities that spans the firm's boundaries, including partners.
• Strategyzer says the Business Model Canvas's nine blocks cover who you serve, what you offer, how you deliver it, what it costs and what it earns.
• Magretta noted that during the dot-com boom "business model" became a buzzword used to justify half-baked plans.
• Six common startup model types: marketplace, platform, B2B2C, open source, hardware and productized service.
• The "who pays" question separates the models most clearly, and misjudging the payer is a common early error.
• Revenue models (commission, subscription, usage, license, one-time sale, advertising, freemium) sit inside a business model and are often combined.
• Johnson, Christensen and Kagermann describe the iPod/iTunes pairing as a nearly $10 billion product in three years, close to half of Apple's revenue, an example of business model innovation.
• A model is a hypothesis to test, not a fixed plan, and it usually changes as a startup matures.

Related Reading

• Marketplace Business Model
• Platform Business Model
• B2B2C Business Model
• Open-Source Business Model
• Hardware Startup Business Model
• Productized Service
• What Are Two-Sided Markets?
• Business Model Canvas
• Unit Economics